Decision engine
EthosGate Decision Outcomes | Governed AI Action
EthosGate decision outcomes are bounded governance signals that describe whether a proposal may be considered, changed, clarified, reviewed, or stopped; they are not action-execution commands.
The five outcomes
ALLOW indicates the proposal appears acceptable within the recorded context. MODIFY indicates a safeguard or change is needed. QUESTION indicates that material information is missing. ESCALATE indicates meaningful uncertainty or conflict requires accountable human review. REFUSE indicates a proposal conflicts with a core safeguard.
Example: a high-impact personnel action
A request to automatically terminate access for a worker based on one unverified signal should not be treated as ALLOW. Missing evidence may lead to QUESTION, potential coercion or downstream harm may lead to ESCALATE, and a request to knowingly deceive the worker may lead to REFUSE.
No production authorization
An outcome does not execute a workflow, sign a transaction, send a message, or override human authority. EthosGate does not claim active production authorization today.